Arxiv:2011.03427V2 [Math.AT] 12 Feb 2021 Riae Fmp Fﬁiest.W Eosrt Hscategor This Demonstrate We Sets

Home , Abuse of notation

arXiv:2011.03427v2 [math.AT] 12 Feb 2021 riae fmp ffiiest.W eosrt hscategor this demonstrate We sets. finite of maps of preimages multiplic sets order-preserving non-commutative an involutive with compatible involution aerlhmlg sn nyteeiopim ntecatego the in epimorphisms the smal only This using homology. homology tahedral hyperoctahedral computes that complex theorem oooyadthe and homology ideal. augmentation ∆ published. never two-si were May’s these to though algebras spaces, group and loop result algebras sketched monoid and of action this mology Fiedor detect groups. did hyperoctahedral construction the bar Hoc of to action the isomorphic detect was sh homology, not had cyclic 6.16] to the g analogously simplicial [FL91, to tion, crossed Loday associated hyperoctahedral and the theory from Fiedorowicz homology constructed the 3]. Section is [FL91, It 2]. tion aerlhmlg oteohrhmlg hoisaiigfr arising theories homology other define the to homology tahedral them. construction compute bar hyperoctahedral to used be can that sma a complexes over chain modules of and product tensor the including homology, the in qiain nnt opspaces. loop infinite equivariant tion, H nScin4w rv htfra umne,ivltv algeb involutive augmented, an for that prove we 4 Section In nScin5w s h eut fScin4t rv connecti a prove to 4 Section of results the use we 5 Section In nScin1w ealtehprcaerlgop.W recall We groups. hyperoctahedral the recall we 1 Section In algebras involutive for homology Hyperoctahedral e od n phrases. and words Key colo ahmtc n ttsis nvriyo Sheffi of University Statistics, 2020 and Mathematics of School nScin3w en yeothda oooyfrinvoluti for homology hyperoctahedral define we 3 Section In yeothda oooyi endi em of terms in defined is homology Hyperoctahedral oooyo ml categories small of homology YEOTHDA OOOYFRIVLTV ALGEBRAS INVOLUTIVE FOR HOMOLOGY HYPEROCTAHEDRAL soitdt h yeothda rse ipiilgrou simplicial crossed hyperoctahedral the to associated yeothda oooywt offiinsi module a in coefficients with homology hyperoctahedral ahmtc ujc Classification. Subject Mathematics opsaebitfo h lsiyn pc ftegroup. the of space inv the classifying under the points from hyperoctahedr fixed built the the space of order, loop homology odd the of to relate group isomorphic is discrete is homology a hyperoctahedral for that theory: construction proves bar paper the hyperoctahedral of the and homology functor rlcosdsmlca ru.I sdfie o nouiea involutive for defined is It group. simplicial crossed dral Abstract. . yeothda oooyi h oooyter associate theory homology the is homology Hyperoctahedral ru oooyo rdc fhprcaerlgroups hyperoctahedral of product a of homology group yeothda oooy rse ipiilgop func group, simplicial crossed homology, hyperoctahedral hc noe h aeifrainb digdt othe to data adding by information same the encodes which , nScin2w ealtencsaybcgon nfunctor on background necessary the recall we 2 Section In . olwn idrwc.W ics h oncino hypero of connection the discuss We Fiedorowicz. following 53,1D3 51,55P47. 55U15, 13D03, 55N35, AILGRAVES DANIEL Introduction 1 a nrdcdb idrwc Fe Sec- [Fie, Fiedorowicz by introduced was uco homology functor yeothda rse ipiilgroup simplicial crossed hyperoctahedral gba vracmuaiern using ring commutative a over lgebras opvaacnrvratbrconstruc- bar contravariant a via roup to.W nrdc the introduce We ation. oncigtehprcaerlho- hyperoctahedral the connecting s wc eosrtdta covariant a that demonstrated owicz oeuvratsal homotopy stable equivariant to d lhmlg ftegopalgebra group the of homology al fFeooiz h anresult main The Fiedorowicz. of shl oooyadteeoedid therefore and homology hschild lto fa qiain infinite equivariant an of olution l,Sefed 37H UK 7RH, S3 Sheffield, eld, lctgr,islf eie functors derived left its category, ll mcosdsmlca rus We groups. simplicial crossed om sioopi o∆ to isomorphic is y e a osrcinadinfinite and construction bar ded n rv a prove and y∆ ry e ope acltshyperoc- calculates complex ler w httehmlg theory homology the that own .Ti aeoyecdsan encodes category This p. eagba ntrso the of terms in algebras ve the nbtenhyperoctahedral between on ateei mle chain smaller a is there ra H otehyperoctahe- the to d yeothda category hyperoctahedral n h lmnso the of elements the and o oooy a construc- bar homology, tor oeie nw as known sometimes , nvra coefficient universal ihcecet in coefficients with H aeoyof category . c- certain modules. We show that when the ground ring is a field of characteristic zero, hyperoctahedral homology of an augmented, involutive algebra can be calculated as the coinvariants of a group action. In Section 6 we show that by appending an initial object to the category ∆H we obtain a symmetric strict monoidal category. We extend the hyperoctahedral bar construction to be defined on this category and demonstrate that we can calculate hyperoctahedral homology using this more structured category. In Section 7 we provide a proof of Fiedorowicz’s theorem for the hyperoctahedral homology of a monoid algebra, for a monoid with involution. This result, Theorem 7.20, relates the hyperoctahedral homology of a monoid algebra to May’s two-sided bar construction and the monads associated to the little intervals and little ∞-cubes operads. In Section 8 we prove the main result of the paper, Theorem 8.8. Consider a discrete group of odd order. Consider its group algebra with involution induced by sending a group element to its inverse. We prove that the hyperoctahedral homology of such a group algebra is isomorphic to the homology of the fixed points under the involution of a C2-equivariant infinite loop space built from the classifying space of the group.

Acknowledgements. I would like to thank Sarah Whitehouse for all her help, support and guidance. I am very grateful to James Brotherston and James Cranch for their helpful sugges- tions.

Notational conventions. Throughout the paper k will denote a unital commutative ring. 2 We let C2 = t | t = 1 denote the cyclic group of order two. For n > 0 we let Σn+1 denote the symmetric group on the set [n] = {0,...,n}. By abuse of notation we will let idn denote both the identity element of Σn+1 and the unique order-preserving bijection [n] → [n]. Unless otherwise stated, homology is taken with coeﬃcients in the ground ring.

1. The Hyperoctahedral Category and Involutive Non-commutative Sets

We recall the deﬁnition of the hyperoctahedral groups Hn+1 for n > 0 and describe the associated category ∆H, following [FL91, Section 3]. We describe the category of involutive non-commutative sets, IF(as) and show that it is isomorphic to ∆H.

Deﬁnition 1.1. For n > 0, the hyperoctahedral group Hn+1 is deﬁned to be the semi-direct n+1 ⋊ n+1 product Hn+1 := C2 Σn+1 where Σn+1 acts on C2 by permuting the factors.

The group Hn+1 is sometimes denoted as a wreath product Σn+1 C2. An element of the hyperoctahedral group Hn+1 is written as a tuple (z0,...,zn; σ) where each zi ∈ C2 and σ ∈ R th Σn+1. Let ti denote the element (1,..., 1, t, 1,..., 1; idn) in Hn+1, where t occurs in the i position for 0 6 i 6 n. Let θj denote the element 1,..., 1; (j j + 1) in Hn+1 for 0 6 j 6 n − 1. Methods for deriving a presentation for a semi-direct product of groups are well-known, see [Joh90, Chapter 10.2] for example. Following these methods one sees that the Hn+1 is generated by the elements of the form ti and θj where the θj satisfy the relations of the symmetric group n+1 Σn+1; the ti satisfy the relations of the group C2 ; the θj and ti commute for i < j and i > j + 1; and θi ◦ ti+1 = ti ◦ θi and θi ◦ ti = ti+1 ◦ θi. The hyperoctahedral groups form a crossed simplicial group [FL91, Section 3]. We recall the associated category ∆H. Recall that the category ∆ has as objects the sets [n] = {0,...,n} for n > 0 and order-preserving maps as morphisms. For n > 0, 0 6 i 6 n + 1 and 0 6 j 6 n, let δi ∈ Hom∆ [n], [n + 1] be the unique order-preserving injection that omits i ∈ [n + 1] and let σ ∈ Hom [n + 1], [n] be the unique order-preserving surjection satisfying σ−1(j) = j ∆ j {j, j + 1}. Recall that these face and degeneracy morphisms generate the category ∆ and are subject to the simplicial relations found in [Lod98, Appendix B], for example. Deﬁnition 1.2. The category ∆H has as objects the sets [n] = {0,...,n} for n > 0. An element of Hom∆H [n], [m] is a pair (ϕ, g) where g ∈ Hn+1 and ϕ ∈ Hom∆ [n], [m] . 2 For (ϕ, g) ∈ Hom∆H [n], [m] and (ψ, h) ∈ Hom∆H [m], [l] the composite is the pair ψ ◦ h (ϕ), ϕ⋆(h) ◦ g ∈ Hom [n], [l] where h (ϕ) and ϕ⋆(h) are determined by the rela- ⋆ ∆H ⋆ tions: (θk)⋆ (δi)= δθk(i), (θk)⋆ (σj)= σθk(j), (tk)⋆ (δi)= δi, (tk)⋆ (σj)= σj and

θk k < i − 1, θk k < j − 1,

⋆ idn k = i − 1, ⋆ θjθj−1 k = j − 1, δi (θk)=  σj (θk)=  idn k = i, θjθj+1 k = j,   θk−1 k > i, θk+1 k > j,       tk k < i, tk k < j, ⋆ ⋆ δi (tk)= idn k = i, σj (tk)= θktk+1tk k = j,   tk−1 k > i, tk+1 k > j,   where the δi and σj are the face and degeneracy maps of the category ∆ and the θk and tk are the generators of the hyperoctahedral group. The hyperoctahedral category encodes the structure of an associative multiplication with a compatible involution. Intuitively speaking, the hyperoctahedral groups encode a total ordering via the symmetric group component and an involution via the cyclic group component whilst the order-preserving maps encode the multiplication. The category of involutive non-commutative sets encodes the same data by adding structure to the preimages of set maps analogously to the category of non-commutative sets of [FT87, A10] and [PR02, 1.2]. Recent work of the author [Gra20] shows that the categories of involutive monoids and involutive bimonoids in a symmetric monoidal category are equivalent to categories of algebras over PROPs constructed from the category of involutive non-commutative sets. We provide an isomorphism between the hyperoctahedral category and the category of involutive non-commutative sets. The category of involutive non-commutative sets ﬁrst appeared in the author’s thesis [Gra19, Part V], where detailed examples and technical checks can be found. We will denote the category of involutive, non-commutative sets by IF(as). It will have as objects the sets [n]= {0,...,n} for n > 0. An element f ∈ HomIF(as) [n], [m] will be a map of sets such that the preimage of each singleton i ∈ [m] is a totally ordered set such that each element comes adorned with a superscript label from the group C2. Remark 1.3. Henceforth we will say that a morphism in IF(as) is a map of sets together with a labelled, ordered set for each preimage. In particular, we will use preimage to mean preimage of a singleton. We will denote composition in IF(as) by • in order to distinguish from the composition of maps of sets. In particular, we use ◦ for two morphisms in IF(as) if we are referring to the composite of the underlying maps of sets.

Definition 1.4. We define an action of C2, which will be denoted by a superscript, on finite, ordered sets with C2-labels by t αj1 αjr tαjr tαj1 j1 < · · · < jr = jr < · · · < j1 . n o n o That is, we invert the ordering and multiply each label by t ∈ C2.

Definition 1.5. Let f1 ∈ HomIF(as) [n], [m] and f2 ∈ HomIF(as) [m], [l] . We define f2 •f1 ∈ Hom [n], [l] to have underlying map of sets f ◦f . We define the labelled totally ordered IF(as) 2 1 set (f • f )−1(i) to be the ordered disjoint union of labelled, ordered sets 2 1 −1 αj f1 (j) . − jαj ∈f 1(i) a2 3 Definition 1.6. The category of involutive, non-commutative sets, IF(as), has as objects the sets [n] = {0,...,n} for n > 0. An element of HomIF(as)([n], [m]) is a map of sets with a total ordering on each preimage such that each element of the domain comes adorned with a superscript label from the group C2. Composition of morphisms is as defined in Definition 1.5. Theorem 1.7. There is an isomorphism of categories ∆H =∼ IF(as). Proof. We note that IF(as) contains the morphisms of ∆. These are the order-preserving maps of sets with the canonical total ordering on each preimage with each label being 1 ∈ C2. Furthermore, IF(as) contains the elements of the hyperoctahedral groups. These are the bijections in IF(as). The isomorphism of categories is similar to the proof of [Gra20, 3.11].

2. Functor homology We recall some constructions from functor homology, or the homology of small categories, from [PR02] and [GZ67]. Let C be a small category. We define the category of left C-modules, denoted CMod, to be the functor category Fun (C, Modk). We define the category of right C-modules, denoted op ModC, to be the functor category Fun (C , Modk). It is well-known that the categories CMod and ModC are abelian with enough projectives and injectives, see for example [PR02, Section 1.6]. Definition 2.1. Let k⋆ denote the right C-module that is constant at the trivial k-module. We will refer to this functor as the k-constant right C-module. Definition 2.2. Let G be an object of ModC and F be an object of CMod. We define the tensor product G ⊗C F to be the k-module

C∈Ob(C) G(C) ⊗k F (C) GL(α)(x) ⊗ y − x ⊗ F (α)(y) where G(α)(x) ⊗ y − x ⊗ F (α)(y ) is the k-submodule generated by the set G(α)(x) ⊗ y − x ⊗ F (α)( y) : α ∈ Hom(C), x ∈ src(G(α)), y ∈ src(F (α)) . This quotient module is spanned k-linearly by equivalence classes of elementary tensors in C∈Ob(C) G(C) ⊗k F (C) which we will denote by [x ⊗ y]. One constructs the tensor product of C-modules as a bifunctor −⊗ −: ModC × CMod → Mod on objects by (G, F ) 7→ G ⊗ F . L C k C Given two natural transformations Θ ∈ HomModC (G, G1) and Ψ ∈ HomCMod (F, F1), the morphism Θ ⊗C Ψ is determined by [x ⊗ y] 7→ ΘC (x) ⊗ ΨC (y) . It is well-known that the bifunctor −⊗C − is right exact with respect to both variables and preserves direct sums, see C for example [PR02, Section 1.6]. We denote the left derived functors of −⊗C − by Tor⋆ (−, −). Recall the nerve of C [Lod98, B.12]. N⋆C is the simplicial set such that NnC for n > 1 consists of all strings of composable morphisms of length n in C and N0C is the set of objects in C. The face maps are defined to either compose adjacent morphisms in the string or truncate the string and the degeneracy maps insert identity morphisms into the string. We will denote an element of NnC by (fn,...,f1) where fi ∈ HomC (Ci−1,Ci). For a small category C and a functor F ∈ CMod there is a simplicial k-module, denoted th C⋆ (C, F ) due to Gabriel and Zisman [GZ67, Appendix 2] whose n homology is canonically C ⋆ isomorphic to Torn (k , F ). Definition 2.3. Let F ∈ CMod. We define

Cn(C, F )= F (C0) (fnM,...,f1) where the sum runs through all elements (fn,...,f1) of NnC and fi ∈ HomC (Ci−1,Ci). We write a generator of Cn(C, F ) in the form (fn,...,f1,x) where (fn,...,f1) ∈ NnC indexes the 4 summand and x ∈ F (C0). The face maps ∂i : Cn(C, F ) → Cn−1(C, F ) are determined by

(fn,...,f2, F (f1)(x)) i = 0, ∂i(fn,...,f1,x)= (fn,...,fi+1 ◦ fi,...,f1,x) 1 6 i 6 n − 1,  (fn−1,...,f1,x) i = n.  The degeneracy maps insert identity maps into the string. By abuse of notation we will also denote the associated chain complex by C⋆ (C, F ). The homology of the associated chain complex will be denoted H⋆ (C, F ). Remark 2.4. This construction is analogous to the Bousﬁeld-Kan construction [BK72, XII.2.1] for the homotopy colimit. This leads to an alternative choice of notation. We can denote the simplicial k-module C⋆ (C, F ) by hocolim F. C C ⋆ This leads to an alternative description of the k-modules Tor⋆ (k , F ) in terms of derived C colimits. One can make the deﬁnition colim (F ) := k ⊗C F following [Lod98, C.10]. One then C denotes the left derived functors by colimi (F ). We will make of use of this alternative notation in Section 5. There is an isomorphic variant of this chain complex constructed using the nerve of the under- category. For details of the under-category see [ML98, II.6] for instance. Let (− ↓ C) : Cop → Cat be the functor that sends an object C ∈ C to the under-category (C ↓ C). For a morphism ′ ′ f ∈ HomC C,C , the functor (f ↓ C) : C ↓ C → (C ↓ C) is determined by precomposition with f. Let k[−]: Set → Modk denote the free k-module functor and, by abuse of notation, its extension to simplicial sets.

Deﬁnition 2.5. Let F ∈ CMod. The chain complex k N⋆ (− ↓ C) ⊗C F has the k-module k N (− ↓ C) ⊗ F in degree n with the boundary map induced from the alternating sum of n C the face maps in the nerve. A generator in degree n is an equivalence class (f ,...,f ) ⊗ x n 0 where f ∈ HomC (C,C ), fi ∈ HomC (Ci− ,Ci) for i > 1 and x ∈ F (C). 0 0 1 The isomorphism of chain complexes k N⋆ (− ↓ C) ⊗C F =∼ C⋆ (C, F ) for each F ∈ CMod is determined by (f ,...,f ) ⊗ x 7→ (f ,...,f , F (f )(x)) on generators in degree n. We will n 0 n 1 0 make use of both variants of this chain complex.

3. Hyperoctahedral Homology We define hyperoctahedral homology for involutive k-algebras in terms of functor homology. We describe the hyperoctahedral homology of the commutative ground ring, which is understood to have the trivial involution. We describe the relationship between hyperoctahedral homology and other homology theories arising from crossed simplicial groups. We prove a universal coefficient theorem for hyperoctahedral homology. Definition 3.1. An involution on an associative k-algebra A is an anti-homomorphism of algebras A → A, which we will denote by a 7→ a, which squares to the identity. A k-algebra equipped with an involution is called involutive. 3.1. Hyperoctahedral bar construction. We define the hyperoctahedral bar construction following Fiedorowicz [Fie, Section 2]. Intuitively, the hyperoctahedral bar construction for an involutive k-algebra A sends the object [n] of ∆H to the tensor power A⊗n+1 and acts on a morphism (ϕ, g) ∈ Hom∆H [n], [m] by sending it to the morphism of k-modules that permutes the factors of the tensor according to the underlying permutation of g, applies the involution according the labels of g and multiplies tensor factors, or inserts identities, according to the order-preserving map ϕ. 5 Definition 3.2. Let A be an involutive, associative k-algebra with the involution denoted by a 7→ a. The hyperoctahedral bar construction is the functor HA : ∆H → Modk given on ⊗n+1 objects by [n] 7→ A . Let (ϕ, g) ∈ Hom∆H [n], [m] . Then ϕ ∈ Hom∆ [n], [m] and g = (z ,...,zn; σ) ∈ Hn . We define 0 +1

< < H zi zi A (ϕ, g) (a0 ⊗···⊗ an)=  ai  ⊗···⊗  ai  i∈(ϕ◦σ)−1(0) i∈(ϕ◦σ)−1(m)  Y   Y  on elementary tensors and extend k-linearly, where the product is ordered according to the map ϕ and a z = 1 azi = i (a zi = t. Note that an empty product is defined to be the multiplicative unit 1A. Remark 3.3. We note that the notation for the hyperoctahedral bar construction presented oct here differs from the notation BA used in Fiedorowicz’s preprint [Fie] and the author’s thesis [Gra19]. We hope the reader will agree that our choice of notation offers equal clarity (H for “hyperoctahedral bar construction” with a subscript A to indicate the algebra under considera- tion) whilst making the typesetting for the remainder of the paper considerably more pleasant. Proofs that the hyperoctahedral bar construction is well-defined can be found in [Fie, Lemma 2.1] or [Gra19, Appendix B]. 3.2. Hyperoctahedral homology. Recall the k-constant right ∆H-module k⋆ from Definition 2.1. Definition 3.4. Let A be an involutive, associative algebra. For n > 0, we define the nth hyperoctahedral homology of A to be ∆H ⋆ H HOn(A) := Torn k , A . Proposition 3.5. The hyperoctahedral homology of the ground ring k, where k has the trivial involution, is isomorphic to k concentrated in degree zero. Proof. The key is to prove that the nerve of the category ∆H is contractible. By [FL91, Example 6], the hyperoctahedral crossed simplicial group {Hn+1} is contractible as a simplicial set with the inclusion maps Hi → Hi+1 forming a contracting homotopy. By [FL91, Proposition 5.8] it follows that ΩB∆H, the based loop space on the classifying space of ∆H, is a contractible space. By considering the long exact sequence in homotopy associated to the path space fibration we deduce that B∆H is a contractible space. By definition the classifying space B∆H is the geometric realization of the nerve N⋆∆H. We deduce that N⋆∆H is contractible and, in particular, its homology is isomorphic to k concentrated in degree zero. The chain complex C⋆ (∆H, Hk), defined following Definition 2.3 is generated k-linearly in degree p by elements of the form fp,...,f1, 1k ⊗···⊗ 1k where fi ∈ Hom∆H [xi−1], [xi] . We observe that there is an isomorphism of chain complexes C (∆H, H ) → k N (∆H) ⋆ k ⋆ determined on generators in degree p by f ,...,f , 1 ⊗···⊗ 1 7→ f ,...,f , from which p 1 k k p 1 the result follows. 3.3. Relationship to dihedral homology. Dihedral homology for algebras with involution was first studied independently by Loday [Lod87] and Krasauskas, Lapin and Solov’ev [KLS87]. It is the homology theory associated to the dihedral crossed simplicial group [FL91, Example 5] and is related to O(2)-equivariant homology [Lod90, Theorem 3.3.3] and Hermitian algebraic K-theory via a dihedral Chern character [KLS87, Section 4]. The category ∆D associated to the dihedral crossed simplicial group is self-dual, that is ∆D =∼ ∆Dop, see for example [Dun89, Proposition 1.4]. Let A be an involutive k-algebra. One observes, similarly to [Fie, Lemma 2.2], that the composite ∼ H op = A D −→ ∆D ֒→ ∆H −−→ Modk∆ 6 is the dihedral bar construction A† of [Lod90, Definition 1.4.1]. Let DA denote the restriction of HA to the subcategory ∆D. Recall the k-constant right C- module from Definition 2.1. By abuse of notation we will use k⋆ to denote both the k-constant right ∆Dop-module and the k-constant right ∆D-module. We observe that ∆Dop ⋆ † ∼ ∆D ⋆ D HD⋆(A) = Tor⋆ k , A = Tor⋆ k , A . It follows that for an involutive k-algebra A, there is a natural map HD⋆ (A) → HO⋆ (A). We can obtain similar maps for symmetric homology and cyclic homology. For an algebra with involution A, the restriction of the hyperoctahedral bar construction HA to the category ∆S sym associated to the symmetric crossed simplicial group is BA , the symmetric bar construction of [Aul10, Definition 13]. We therefore have a natural map HS⋆ (A) → HO⋆ (A), constructed similarly to the dihedral case. Pre-composing with the natural map HC⋆ (A) → HS⋆ (A) of [Aul10, Section 12] yields, for an involutive k-algebra, a natural map HC⋆ (A) → HO⋆ (A). 3.4. Universal coefficient theorem. We show that hyperoctahedral homology satisfies a universal coefficient theorem analogous to that satisfied by symmetric homology [Aul10, Theorem 21]. Definition 3.6. Let A be an involutive k-algebra and let M be a k-module. Let d denote the boundary map of C⋆ (∆H, HA). We define the hyperoctahedral homology of A with coefficients in M, denoted HO⋆ (A; M), to be the homology of the chain complex C⋆ (∆H, HA) ⊗ M with boundary map d ⊗ idM .

Let A be an involutive k-algebra. One observes that for each n > 0, HOn (A; −) is a covariant endofunctor on the category of k-modules. Explicitly, let M be a k-module thought of as a chain complex concentrated in degree zero. The functor HOn (A; −) is defined by first taking th the tensor product of chain complexes C⋆ (∆H, HA) ⊗− and then taking the n homology. Since the tensor product of k-modules and taking homology both commute with direct sums we see that the functors HOn (A; −) are additive. A standard argument shows that if A is flat over k, given a short exact sequence of k-modules, the functors HOn (A; −) give rise to a long exact sequence in homology. Theorem 3.7. Let A be a flat, involutive k-algebra and let M be a k-module. There is a spectral sequence with p,q ∼ k E2 = Torp HOq (A) , M converging to HO⋆ (A; M). Proof. This is the “universal coefficient spectral sequence” of [McC01, Theorem 12.11]. We note that the boundedness condition is satisfied since HOq (A; −)=0 for q < 0. The tensor product of k-modules commutes with arbitrary direct sums as does taking homology and so the functors HOq (A; −) commute with arbitrary direct sums. We recall some definitions from homological algebra [Wei94, 4.1, 4.2]. Recall that the weak k global dimension of a commutative ring k is the largest integer n such that Torn (M,N) is non-zero for some k-modules M and N. We will denote it wgd(k). Furthermore, recall that a commutative ring k is said to be hereditary if every ideal is projective. Corollary 3.8. If the ground ring k satisfies wgd(k) 6 1 then the spectral sequence of Theorem 3.7 reduces to short exact sequences k 0 → HOn(A) ⊗ M → HOn (A; M) → Tor1 HOn−1 (A) , M → 0. Furthermore, if k is hereditary and A is projective these short exact sequences split, although not naturally. k Proof. By definition, if wgd(k) 6 1 then Torp HOq (A) , M = 0 for all p > 1. The corollary now follows from [Dol62, Korollar 2.13]. If k is hereditary and A is projective the splitting follows from [Rot09, Corollary 7.56]. 7 4. Hyperoctahedral Homology of an Augmented Algebra We prove that in the case of an augmented involutive k-algebra there exists a smaller chain complex that computes hyperoctahedral homology. We will do this in two stages. We will identify a subcomplex of the standard complex which can be used to calculate the hyperoctahedral homology of an augmented, involutive k-algebra up to a copy of k in degree zero. This subcomplex contains only those elements of A that lie in the augmentation ideal. We will then introduce the epimorphism construction for hyperoctahedral homology. Using this construction we reduce the complex further, showing that we need only use the epimorphisms in ∆H to calculate the hyperoctahedral homology. The material in this section first appeared in the author’s thesis [Gra19, Part VI] and was inspired by the analogous construction for symmetric homology [Aul10, Section 6]. Throughout this section we will denote an element of Np [x] ↓ ∆H by fp,...,f1,f0 . In particular, f0 = (ϕ, g) ∈ Hom∆H [x], [x0] and f = (ϕ , g ) ∈ Hom [x ], [x ] for i > 1. i i i ∆H i−1 i 4.1. Reduced hyperoctahedral homology. An associative k-algebra A is said to be augmented if it is equipped with a k-algebra homomorphism ε: A → k. The augmentation ideal I of A is defined to be Ker(ε). Note that the structure of an augmented algebra is compatible with an involution. In particular, the augmentation ideal is closed under involution. Recall from [LV12, Section 1.1.1] that for an augmented k-algebra there is an isomorphism of k-modules A =∼ I ⊕ k. It follows that every element a ∈ A can be written uniquely in the form y + λ where y ∈ I and λ ∈ k. Definition 4.1. Let A be an augmented k-algebra with augmentation ideal I. A basic tensor ⊗n in A is an elementary tensor a1 ⊗···⊗an such that either ai ∈ I or ai = 1k for each 1 6 i 6 n. A tensor factor ai is called trivial if ai = 1k and is called non-trivial if ai ∈ I. Note that for an augmented k-algebra A, the k-module A⊗n is generated by all finite k-linear combinations of its basic tensors. Henceforth in this section A will denote an augmented, involutive k-algebra with augmentation ideal I. Consider the chain complex k N⋆ (− ↓ ∆H) ⊗∆H HA defined analogously to Definition 2.5, whose homology is the hyperoctahedral homology of A. By analysing the quotient module one can show that k N⋆ (− ↓ ∆H) ⊗∆H HA is generated k-linearly in degree n by those equivalence classes of the form (f ,...,f ,f ) ⊗ Y such that Y is a basic tensor n 1 0 x x containing no trivial factors and those equivalence classes of the form (f ,...,f ,f ) ⊗ 1 . n 1 0 k Note that for this last set of equivalence classes the domain of f is the set [0]. We observe that 0 the boundary map of k N⋆ (− ↓ ∆H) ⊗∆H HA preserves these two types of equivalence classes. We can therefore make the following definition. Definition 4.2. Let C⋆ (∆H,I) denote the subcomplex of k N⋆ (− ↓ ∆H) ⊗∆H HA generated k-linearly in degree n by all equivalence classes (f ,...,f ,f ) ⊗ Y such that Y is a basic n 1 0 x x tensor containing no trivial factors. We denote the homology of this subcomplex by H⋆ (∆H,I). Let C⋆ (∆H, k) denote the subcomplex of k N⋆ (− ↓ ∆H) ⊗∆H HA generated k-linearly in degree n by all equivalence classes of the form (f ,...,f ,f ) ⊗ 1 . n 1 0 k Proposition 4.3. Let A be an involutive augmented k-algebra with augmentation ideal I. For each n > 1 there is an isomorphism of k-modules HOn (A) =∼ Hn (∆H,I). Furthermore, there is an isomorphism of k-modules HO0 (A) =∼ H0 (∆H,I) ⊕ k.

Proof. The inclusion of the subcomplex C⋆ (∆H,I) into k N⋆ (− ↓ ∆H) ⊗∆H HA has a left inverse given by the canonical projection map, the kernel of which is C⋆ (∆H, k). This gives a splitting of k N⋆ (− ↓ ∆H) ⊗∆H HA as a direct sum C⋆ (∆H,I) ⊕ C⋆ (∆H, k). In order to prove the result we will invoke [Wei94, 8.4.6] to show that the homology of C⋆ (∆H, k) is isomorphic to k concentrated in degree zero. One defines an augmentation ε: C0 (∆H, k) → k by [f ⊗ 1k] 7→ 1k. One then checks that the maps hn : Cn (∆H, k) → Cn+1 (∆H, k) determined by (fn,...,f1,f) ⊗ 1k 7→ (fn,...,f1, f, id0) ⊗ 1k for n > 0 and h−1 : k → C0 (∆H, k) determined by 1k 7→ [id0 ⊗ 1k] satisfy the conditions described in [Wei94, 8.4.6]. 8 Definition 4.4. Let A be an augmented, involutive k-algebra. For n > 0, we define the nth reduced hyperoctahedral homology of A to be

HOn(A) := Hn (∆H,I) . 4.2. Epimorphism construction. We begin by noting that a morphism (ϕ, g) in the category ∆H is an epimorphism if and only ifg the underlying map of sets is surjective.

Remark 4.5. Consider a morphism (ϕ, g) ∈ Hom∆H [x], [z] and let r = Im (ϕ) . By the unique decomposition of morphisms in ∆ [Lod98, Theorem B.2] the morphism (ϕ, g) can be written uniquely in the form iϕ, idr−1 ◦ πϕ, g where πϕ, g ∈ HomEpi∆H [x], [r − 1] and iϕ : [r − 1] → [z] is a monomorphism in the category ∆. Whenever we use this decomposition we will abuse notation and write iϕ for iϕ, idr−1 ∈ Hom∆H [r − 1], [z] . Definition 4.6. Let Epi∆H denote the subcategory of ∆H whose morphisms are the epimorphisms in ∆H. Definition 4.7. Let A be an augmented, involutive k-algebra with augmentation ideal I. We ⊗n+1 define a functor HI : Epi∆H → Modk on objects by [n] → I . On morphisms we define HI to be the restriction of HA to the subcategory Epi∆H.

Remark 4.8. The functor HI is well-deﬁned since we are only taking the epimorphisms of ∆H and the augmentation ideal I is closed under the involution and under multiplication.

We deﬁne the chain complex k N⋆ (− ↓ Epi∆H) ⊗Epi∆H HI following Deﬁnition 2.5. We note that there is an inclusion of chain complexes i ⋆ : k N⋆ (− ↓ Epi∆H) ⊗Epi∆H HI → C⋆ (∆H,I) . We claim that this inclusion is a chain homotopy equivalence. In order to prove this we introduce the epimorphism construction for hyperoctahedral homology. Lemma 4.9. A morphism

(ϕ1,g1) (ϕ2,g2) (ψ, h) : [x] −−−−→ [z1] → [x] −−−−→ [z2] in [x] ↓ ∆H induces a unique morphism

(π ,g ) (π ,g ) ϕ1 1 ϕ2 2 (ψ, h) : [x] −−−−−→ [r1 − 1] → [x] −−−−−→ [r2 − 1] in [x] ↓ Epi∆H . Proof. Using Remark 4.5, we have a commutative diagram

(ψ,h)

(ϕ1,g1) (ϕ2,g2) [z1][x][z2]

iϕ1 (πϕ1 ,g1) (πϕ2 ,g2) iϕ2

[r1 − 1] [r2 − 1] in ∆H.

By construction, Im iϕ1 = Im (ϕ1, g1) . From the commutativity of the diagram one deduces that Im (ψ, h) ◦ i = Im i . ϕ1 ϕ2 By Remark 4.5 there exists a unique morphism (ψ, h) : [r − 1] → [r − 1] in Epi∆H such 1 2 that (ψ, h) ◦ iϕ1 = iϕ2 ◦ (ψ, h).

By commutativity of the diagram we see that iϕ2 ◦(ψ, h)◦(πϕ1 , g1)= iϕ2 ◦(πϕ2 , g2) and since iϕ2 is a monomorphism in ∆H we can cancel on the left and deduce that (ψ, h)◦(πϕ1 , g1) = (πϕ2 , g2) as required. 9 A routine check, which can be found in [Gra19, Proposition 27.8.5], demonstrates that this assignment on morphisms is functorial. We can therefore make the following deﬁnition.

Definition 4.10. For x > 0, we define the functor Ex : [x] ↓ ∆H → [x] ↓ Epi∆H on objects by (ϕ, g) 7→ π , g and on morphisms by (ψ, h) 7→ (ψ, h). We call E the epimorphism ϕ x construction for hyperoctahedral homology. We observe that by the functoriality of the nerve construction, the functor Ex induces a map of simplicial sets N⋆Ex : N⋆([x] ↓ ∆H) → N⋆([x] ↓ Epi∆H). This induced map takes a string of composable morphisms in ∆H and sends it to the string of composable morphisms in Epi∆H obtained by applying the epimorphism construction. Definition 4.11. We define a map of chain complexes

χ⋆ : C⋆ (∆H,I) → k N⋆(− ↓ Epi∆H) ⊗Epi∆H HI to be determined in degree n by (fn,...,f1,f0) ⊗ Yx 7→ k [NnEx] (fn,...,f1,f0) ⊗ Yx .

Remark 4.12. One must check that the maps χn are well-deﬁned. This consists of a routine but lengthy check that the epimorphism construction is compatible with equivalence classes under the tensor products −⊗∆H − and −⊗Epi∆H −. The full details can be found in [Gra19, Proposition 27.9.1]. Theorem 4.13. Let A be an augmented, involutive k-algebra with augmentation ideal I. For n > 0 there exist isomorphisms of k-modules

HOn(A) =∼ Hn k N⋆(− ↓ Epi∆H) ⊗Epi∆H HI =∼ Hn C⋆ (Epi∆H, HI ) . Proof. Weg show that the inclusion map of chain complexes i⋆ : k N⋆ (− ↓ Epi∆H) ⊗Epi∆H HI → C⋆ (∆H,I) is a chain homotopy equivalence with inverse given by χ⋆. By construction, the composite χ⋆ ◦i⋆ is equal to the identity map on the chain complex k N⋆ (− ↓ Epi∆H) ⊗Epi∆H HI . One shows that the composite i⋆ ◦ χ⋆ is homotopic to the identity map on the chain complex C⋆ (∆H,I) by constructing a presimplicial homotopy between the two maps. We will make use of Remark 4.5. n Let hj : Cn (∆H,I) → Cn+1 (∆H,I) be determined by

(fn,...,f1,f0) ⊗ Yx 7→ fn,...,fj+1, iϕj , fj,..., f1, Ex(f0) ⊗ Yx . n That is, hj is the map determined by mapping an equivalence class in Cn (∆H,I) to the equivalence class in Cn+1 (∆H,I) obtained by applying the epimorphism construction to the

ﬁrst j morphisms in the representative, inserting iϕj from the unique decomposition of Remark 4.5 and leaving the remaining morphisms unchanged. A routine check, which can be found in [Gra19, Proposition 27.10.2], demonstrates that these maps do indeed form a presimplicial homotopy between i⋆ ◦ χ⋆ and the identity map.

5. Hyperoctahedral Homology and Group Homology We provide an application of our smaller complex for computing the hyperoctahedral homology of an augmented, involutive algebra A. We will show that in this case we can express the reduced hyperoctahedral homology of A in terms of the group homology of a product of hyperoctahedral groups with coefficients in modules constructed from the epimorphisms in ∆H. When the ground ring is a field of characteristic zero one can say more: the reduced hyperoctahedral homology of A can be computed as the coinvariants of a group action. The key to this construction is S lomińska’s work on E-I-A-categories [S l91], that is categories for which all endomorphisms are isomorphisms and all isomorphisms are automorphisms. We observe that 10 Epi∆H is just such a category. The results in this section are analogous to the symmetric homology case [Aul10, Section 7]. The notation throughout this section is chosen to closely resemble [Sl91] and [Aul10, Section 7].

Deﬁnition 5.1. Let S0 denote the opposite category of the poset of non-empty ﬁnite subsets of N. The set of objects are non-empty ordered tuples of non-negative integers (xr < · · ·

Deﬁnition 5.2. We deﬁne a functor A: S0 → Grp, where Grp is the category of groups, on objects by r r

A (xr < · · ·

Deﬁnition 5.3. We deﬁne a functor E : S0 → Set on objects by

{x0} r = 0 E (xr < · · · 1, where {x0} is the one element set containingQ x0. For a morphism f : (xr < · · ·

Deﬁnition 5.4. Let (xr < · · ·

−1 −1 (gn,...,g0) × (fn,...,f1) 7→ gnfngn−1,...,g1f1g0 . Remark 5.5. We observe that this action is compatible with morphisms in S0. That is, for ′ ′ a morphism f : (xr < · · ·

EI (xr < · · ·

Remark 5.7. Observe that k E (xr < · · · 0. We can express these group homology functors in terms of the homology of small categories. Recall the notation of Remark 2.4. We ﬁx X = (xr < · · ·

p,q S0 E2 = colimp Hq A(−), EI (−) ⇒ HOp+q(A). 11 g Proof. Recall the notation of Remark 2.4 and consider

Y = hocolim hocolim EI. S0 GX This is a bisimplicial k-module with the horizontal simplicial structure induced from the nerve of GX and the vertical simplicial structure induced from the nerve of S0. Taking the “horizontal homology ﬁrst” spectral sequence associated to this bisimplicial k-module we obtain a spectral sequence of the form p,q S0 GX E2 = colimp colimq EI ⇒ Hp+q(Y ). By [Sl91, Proposition 1.6] and Theorem 4.13 we have ∼ Epi∆H H Hp+q(Y ) = colimp+q I = HOp+q(A) as required. g Corollary 5.9. Suppose that k is a ﬁeld of characteristic zero. Then

S0 HOp(A) = colimp EI (X)A(X), where EI (X)A(X) denotes the coinvariants of EI (X) under the action of A(X). g k[G] Proof. Recall that H⋆ (G, M) = Tor⋆ (k, M) for a group G and a left G-module M [Lod98, C.3]. When k is a field of characteristic zero and G is finite, M is projective [Bro82, I.8 Exercise 5] and so Hq A(X), EI (X) = 0 for all q > 0. Therefore the spectral sequence of Theorem 5.8 collapses as required. 6. An Extension of the Hyperoctahedral Category We extend the hyperoctahedral category ∆H by appending an initial object. This extra data yields the structure of a symmetric strict monoidal category under the disjoint union. We can extend the hyperoctahedral bar construction to this new category. One benefit of this new formulation will become apparent when we prove Fiedorowicz’s theorem for the hyperoctahedral homology of monoid algebras in Section 7.

Deﬁnition 6.1. Let ∆H+ be the category formed by appending an initial object, the empty set ∅, to the category ∆H. For n > 0 we will denote the unique morphism ∅→ [n] by in.

Proposition 6.2. ∆H+ is a symmetric strict monoidal category under the disjoint union of sets.

Proof. On objects we observe that [n] ∐ [m] = [n + m + 1]. For (ϕ, g) ∈ Hom∆H+ [n], [n1] and (ψ, h) ∈ Hom [m], [m ] the morphism ∆H+ 1 (ϕ, g) ∐ (ψ, h): [n + m + 1] → [n1 + m1 + 1] is defined by (ϕ, g) acting on the first n + 1 elements and (ψ, h) acting on the remaining m + 1 points. One can check that the disjoint union is associative with the empty set as a unit. The transposition functor [n]∐[m] → [m]∐[n] is the identity on objects and on morphisms is defined by precomposition with the block permutation that transposes the first n + 1 elements with the remaining m + 1 elements.

We can extend the hyperoctahedral bar construction to the category ∆H+.

Definition 6.3. Let HA+ : ∆H+ → Modk be the functor defined as follows. On the subcategory ⊗n+1 ∆H we define HA+ = HA. We define HA+ ∅ = k. We define HA+(in): k → A to be the inclusion of k-algebras. Theorem 6.4. Let A be an involutive k-algebra. There is an isomorphism of graded k-modules HO⋆(A) =∼ H⋆ (∆H+, HA+). Proof. One can easily check that the arguments of [Aul10, Section 4] can be applied in the hyperoctahedral case. 12 7. Fiedorowicz’s Theorem for Monoid Algebras In this section we provide a proof of Fiedorowicz’s theorem for the hyperoctahedral homology of monoid algebras, Theorem 7.20. This result originally appeared in [Fie, Proposition 2.3] but was never published. It relates the hyperoctahedral homology of a monoid algebra for a monoid with involution to the homology of May’s two-sided bar construction. Throughout this section all topological spaces are assumed to be compactly generated weak Hausdorff. We will assume that the category of based topological spaces and the category of based simplicial sets are equipped with the classical Quillen model structures. In particular, a weak homotopy equivalence is a map that induces isomorphisms on homotopy groups. Furthermore, a C2-weak homotopy equivalence is a C2-equivariant map which induces weak homotopy equivalences on the fixed point spaces for both subgroups of C2.

7.1. Categories and monoids. Let C2-Set⋆ denote the category of finite based sets with a basepoint-preserving C2-action. The morphisms are basepoint-preserving, action-preserving maps of sets. Let C2-Top⋆ denote the category of based topological spaces with a basepoint- preserving C2-action. The morphisms are basepoint-preserving continuous maps compatible with the C2-action. Definition 7.1. Let G be a discrete group. Let EG be the category whose objects are the elements of G with a unique morphism from each object to any other. Let E⋆G denote the nerve of this category. Let EG denote the geometric realization of E⋆G. Let EG+ denote the space EG with a disjoint basepoint appended. We will refer to both E⋆G and EG as the total space of G. Definition 7.2. A monoid with involution is a monoid M, for which the unit element is a non-degenerate basepoint, together with an anti-homomorphism of monoids M → M, which will be denoted by m 7→ m. A topological monoid with involution is a monoid with involution equipped with a topology for which the binary operation and the involution are continuous. Remark 7.3. Note that a monoid with involution can be thought of as a topological monoid with involution by equipping it with the discrete topology. 7.2. The two-sided bar construction. We recall the two-sided bar construction from [May72, Section 9]. A monad in the category Top⋆ is an endofunctor F together with natural transformations µ: F F ⇒ F and η : 1 ⇒ F , where 1 is the identity functor on Top⋆, satisfying the conditions of [May72, Definition 2.1]. An F -algebra is a based topological space X together with a map ξ : F X → X, compatible with µ and η in the sense of [May72, Definition 2.2]. An F -functor is a functor G, whose source category is Top⋆, together with a natural transformation λ: GF ⇒ G, compatible with µ and η in the sense of [May72, Definition 9.4]. Let F be a monad in Top⋆. Let X be an F -algebra and let G be an F -functor. The two- sided bar construction B⋆ (G, F, X), is the simplicial space defined in degree n by Bn (G, F, X)= GF nX, with the 0-face being induced by the natural transformation λ, the n-face induced by the map ξ and the intermediate faces being induced by the natural transformation µ. The degeneracies are induced by the natural transformation η. We denote the based topological space obtained by taking the geometric realization by B (G, F, X).

7.3. A∞-operads. We recall some facts about A∞-operads in Top⋆, in particular the trivial operad and the little intervals operad, following [Fie84, 7.4]. Note that we are using the non- symmetric form of A∞-operad, see [May72, 3.12, 3.14]. An A∞-operad O is a collection of contractible spaces O(n) : n > 0 , such that O(0) = ⋆, there is a unit element 1 ∈ O(1) and there exist composition maps γ : O(n) × O(k1) ×···×O(kn) → O(k1 + · · · + kn) satisfying Conditions 1 and 2 of [May72, Deﬁnition 1.1]. 13 An O-space is a based topological space X together with maps θ : O(n) × Xn → X satisfying Conditions 1 and 2 of [May72, Lemma 1.4]. The monad associated to an A∞-operad, O, is the functor O : Top⋆ → Top⋆ deﬁned on objects by n > O(n) × X O(X)= n 0 ` ≈ where ≈ denotes the subspace generated by the equivalence relation induced from the insertion and deletion of basepoints. That is, it is the functor constructed from O via [May72, Construction 2.4] where we omit the equivariance conditions.

Example 7.4. The first example of an A∞-operad is the trivial operad, J = {⋆: n > 0}. A J -space is a topological monoid. The associated monad J is the James construction, originally n defined in [Jam55, Section 1]. For X ∈ Top⋆ with basepoint e, J(X) is the quotient of n>1 X by the equivalence relation generated by all identifications of the form ` (x1,...,xi−1,e,xi+1,...,xn) ∼ (x1,...xi−1,xi+1,...,xn)

For a morphism f : X → Y in Top⋆, J(f) is the morphism deﬁned by applying f point-wise.

Example 7.5. The second example of an A∞-operad is the little intervals operad, C1, originally due to Boardman and Vogt [BV68, Example 5]. The space

C1(n)= [x1,y1],..., [xn,yn] : 0 6 x1

γ : C1(n) ×C1(k1) ×···×C1(kn) →C1(k1 + · · · + kn)

th are deﬁned by embedding the intervals of C1(ki) into the i interval of an element of C1(n). The monad C1 associated to C1 is called the little intervals monad.

We can extend the James construction, J, and the little intervals monad, C1, to functors on the category C2-Top⋆ following [Ryb91] and [Xic06, Section 3] respectively.

Deﬁnition 7.6. We extend the James construction to a functor J : C2-Top⋆ → C2-Top⋆ by deﬁning the C2-action t (x1,...,xn) = (txn,...,tx1).

th Definition 7.7. We define a C2-action on C1(n), the n space of the little intervals operad, by reflecting the unit interval about its midpoint. That is,

t [x1,y1],..., [xn,yn] = [1 − yn, 1 − xn],..., [1 − y1, 1 − x1] . C We extend the little intervals monad to a functor 1 : C2-Top⋆ → C2-Top⋆ by deﬁning the C2-action

t [x1,y1],..., [xn,yn] , (x1,...,xn) = t [x1,y1],..., [xn,yn] , (txn,...,tx1) . h i h i Proposition 7.8. There is a natural C2-weak homotopy equivalence C1(X) → J(X) for any X ∈ C2-Top⋆.

Proof. For any A∞-operad, O, there is a unique map of operads O→J into the trivial operad, which induces a map of monads O → J such that O(X) is naturally weakly homotopy equivalent to J(X). One observes that the map of monads C1 → J induced by the unique map of A∞- operads C1 →J , is compatible with the C2-actions of Definitions 7.6 and 7.7 and the C2-weak homotopy equivalence follows. 14 7.4. Hyperoctahedral operad. An E∞-operad O consists of a collection of contractible spaces O(n) : n > 0 , such that there is a free right action of the symmetric group Σn on O(n), O(0) = ⋆, there is a unit element 1 ∈ O(1) and there exist composition maps γ : O(n) × O(k1) ×···×O(kn) → O(k1 + · · · + kn) satisfying Conditions 1–3 of [May72, Definition 1.1]. An O-space is a based topological space X together with maps θ : O(n) × Xn → X satisfying Conditions 1–3 of [May72, Lemma 1.4]. The monad associated to an E∞-operad, O, is the functor O : Top⋆ → Top⋆ defined on objects by n > O(n) ×Σn X O(X)= n 0 ` ≈ where ≈ denotes the subspace generated by the equivalence relation induced from the insertion and deletion of basepoints. That is, it is the functor constructed from O via [May72, Construction 2.4]. One example of an E∞-operad is the Barratt-Eccles operad [BE74] (see also [BF04, 1.1]). th The n space of the Barratt-Eccles operad is EΣn, the total space of the symmetric group Σn. We define the analogous E∞-operad where the symmetric groups are replaced with the hyperoctahedral groups.

Deﬁnition 7.9. Let H denote the hyperoctahedral operad. For n > 1, let H(n) = EHn, the total space of Hn. Analogously to the Barratt-Eccles operad, the composition operations are determined by the natural maps

Hr × Hn1 ×···× Hnr → Hn1+···+nr where the elements of the Hni act in blocks and the element of Hr acts on the blocks. The right action of Σn on H(n) is the diagonal action.

Deﬁnition 7.10. Let Z : Top⋆ → Top⋆ denote the monad associated to the hyperoctahedral operad H.

Remark 7.11. Since H is an E∞-operad, there exists a natural weak homotopy equivalence C C Z(X) ≃ ∞(X) for each X ∈ Top⋆, where ∞ is the monad associated to the little ∞-cubes operad of [May72, Section 4]. 7.5. Hyperoctahedral bar construction for monoids.

Definition 7.12. Let M be a monoid with involution. We define a functor HM : ∆H+ → C2-Set⋆ on objects by M n+1 n > 0 HM [n] = (∅ n = −1. n+1 where M denotes the (n + 1)-fold Cartesian product. HM is defined on morphisms in ∆H n+1 analogously to HA. We define HM (in) to be the unique map ∅ → M . We call HM the hyperoctahedral bar construction for monoids. Lemma 7.13. Let M be a monoid with involution. The hyperoctahedral homology of the monoid H algebra k[M] is isomorphic to the homology of the simplicial set N⋆ (− ↓ ∆H+) ×∆H+ M . Proof. Recall that the free k-module functor sends coproducts to direct sums and products to tensor products. One observes that applying the free k-module functor to the simplicial set H H N⋆ (− ↓ ∆H+) ×∆H+ M we obtain the chain complex k N⋆ (− ↓ ∆H+) ⊗∆H+ k[M] whose homology calculates the hyperoctahedral homology of the monoid algebra k[M]. Remark 7.14. By considering M equipped with the discrete topology we note that the functor H H M lands in the category C2-Top⋆ and N⋆ (− ↓ ∆H+) ×∆H+ M is a simplicial space whose geometric realization is homeomorphic to the homotopy colimit of HM . That is,

HO⋆ k[M] =∼ H⋆ hocolim HM . ∆H + 15 For the rest of this section all homotopy colimits will be taken over the category ∆H+. In order to ease the typesetting we will henceforth omit the category from the notation. 7.6. Technical lemmata. Fiedorowicz’s theorem, Theorem 7.20, will relate the hyperoctahedral homology of an involutive monoid algebra of a discrete monoid to the homology of the two-sided bar construction, involving the C2-equivariant little intervals monad C1 and the monad C∞ associated to the little ∞-cubes operad. The method of proof involves replacing an involutive monoid with a resolution in terms of the bar construction and the C2-equivariant James construction. One applies the hyperoctahedral bar construction and the homotopy colimit and analyses the result. In order to prove this theorem we must ﬁrst analyse the homotopy colimit of the functor H J(X) : ∆H+ → C2-Top⋆ for a discrete based C2-space. The homotopy colimit is the geometric H realization of the simplicial space N⋆ (− ↓ ∆H+)×∆H+ J(X). We will show that this homotopy colimit can be expressed in terms of the monad Z associated to the hyperoctahedral operad. For ease of indexing we will denote the empty set by [−1] throughout this subsection. For a based topological space X, we deﬁne X0 to be the basepoint of X. H The following lemma shows that we can express the homotopy colimit of the functor J(X) in terms of the underlying space X.

Lemma 7.15. Let X ∈ C2-Top⋆ have the discrete topology. There is an isomorphism of simplicial spaces N [n] ↓ ∆H × Xn+1 H ∼ n>−1 ⋆ + Hn+1 N⋆ (− ↓ ∆H+) ×∆H+ J(X) = ` ≈ where ≈ is the subspace generated by the equivalence relation induced from the insertion and deletion of basepoints. Proof. The technical content of this proof is analogous to [Aul10, Lemma 33]. We will motivate the proof here. By deﬁnition, the left hand side is equal to n+1 n>−1 N⋆ [n] ↓ ∆H+ × J(X) ` G(α)(x) ⊗ y − x ⊗F (α)(y) where G(α)(x) ⊗ y − x ⊗ F (α)( y) is the subspace generated by the relations on ∆H+ morphisms, analogously to Deﬁnition 2.2.

A morphism in ∆H+ can be written uniquely as a composite δ ◦ s ◦ g, where g is an auto- morphism, that is an element of the hyperoctahedral group, s is a surjective order-preserving map and δ is an injective order-preserving map. Taking the quotient of n+1 N⋆ [n] ↓ ∆H+ × J(X) n>−1 a by the subspace generated by the relation for automorphisms we obtain n+1 V = N⋆ [n] ↓ ∆H+ ×Hn+1 J(X) . n>−1 a Extending the argument of [Aul10, Lemma 33] to the hyperoctahedral case one can show that any element of the quotient space of V by the subspace generated by the relation for order- preserving surjections can be expressed uniquely as an element of n+1 N⋆ [n] ↓ ∆H+ ×Hn+1 X . n>−1 a Finally, one observes that the subspace generated by the relation for injective order-preserving maps is precisely the subspace generated by the relation induced from the insertion and deletion of basepoints.

Lemma 7.16. There exist weak homotopy equivalences of simplicial sets E⋆Σn+1 → E⋆Hn+1 and E⋆Hn+1 → N⋆ [n] ↓ ∆H+ . 16 Proof. All three simplicial sets are connected and contractible. The ﬁrst weak homotopy equivalence is determined by the inclusion of the symmetric group into the hyperoctahedral group. The second is determined by

−1 −1 −1 −1 g1g0 ,...,gngn−1 7→ g0, g1g0 ,...,gngn−1 in degree n. H The following lemma demonstrates that the homotopy colimit of the functor J(X) is weakly homotopy equivalent to the monad Z of the hyperoctahedral operad applied to the based topological space EC2+ ∧C2 X.

Lemma 7.17. Let X ∈ C2-Top⋆ have the discrete topology. There is a natural weak homotopy equivalence of simplicial spaces

n+1 n+1 > N⋆ [n] ↓ ∆H+ ×Hn+1 X > E⋆Hn+1 ×Σn+1 EC2+ ∧C2 X n −1 ≃ n −1 ` ≈ ` ≈ where ≈ is the subspace generated by the equivalence relation induced from the insertion and deletion of basepoints. Proof. Let n+1 S = N⋆ [n] ↓ ∆H+ ×Hn+1 X . n>−1 a The total space EC2+ is a contractible space with a free C2-action and so the projection map EC2+ ∧ X → X is a weak homotopy equivalence. That is, we can replace X with a free C2-space. Furthermore, by Lemma 7.16, we can replace the nerve of the under-category with the total spaces of the hyperoctahedral groups. We obtain weak homotopy equivalences n+1 n+1 S ≃ N⋆ [n] ↓ ∆H+ ×Hn+1 (EC2+ ∧ X) ≃ E⋆Hn+1 ×Hn+1 (EC2+ ∧ X) . n>−1 n>−1 a a n+1 We observe that the action of the hyperoctahedral group Hn+1 on (EC2+ ∧ X) is free. Let (z0,...,zn; σ) ∈ Hn+1. This group element acts ﬁrstly by applying the free C2-action point-wise in the product, according to the elements z0,...,zn ∈ C2. It then freely permutes the factors of the product according to the permutation σ ∈ Σn+1. By the properties of free group actions we observe that n+1 n+1 (EC2+ ∧ X) (EC2+ ∧C2 X) n+1 S ≃ ≃ ≃ E⋆Σn+1 ×Σn+1 (EC2+ ∧C2 X) . Hn Σn n>− +1 n>− +1 n>− a1 a1 a1 By Lemma 7.16 we can replace the total spaces of the symmetric groups with the total spaces of the hyperoctahedral groups: n+1 S ≃ E⋆Hn+1 ×Σn+1 EC2+ ∧C2 X . n>−1 a A routine check shows that these weak homotopy equivalences are compatible with taking the quotient by the subspace generated by the equivalence relation induced from the insertion and deletion of basepoints.

Corollary 7.18. Let X ∈ C2-Top⋆ have the discrete topology. There are natural weak homotopy equivalences of based topological spaces H C hocolim J(X) ≃ Z EC2+ ∧C2 X ≃ ∞ EC2+ ∧C2 X where Z is the monad associated to the hyperoctahedral operad and C∞ is the monad associated to the little ∞-cubes operad. Proof. The ﬁrst weak homotopy equivalence follows from Lemma 7.17 upon taking geometric realization. The second follows from Remark 7.11. 17 Lemma 7.19. Let M be a monoid with involution. The weak homotopy equivalence of Corollary 7.18 induces a weak homotopy equivalence H B⋆ hocolim J(−), J, M → B⋆ Z EC2+ ∧C2 − , J, M . Proof. One observes that the only simplicial map in the bar construc tion that the equivalence interacts with is the 0-face. Since the equivalence of Corollary 7.18 is natural we have a com- muting square

H ≃ hocolim J(J(M)) Z EC2+ ∧C2 J(M)

∂0 ∂0 H ≃ hocolim J(M) Z EC2+ ∧C2 M induced by the map J(M) → M as required.

7.7. Fiedorowicz’s theorem for monoid algebras. For ease of notation for the remainder E of the paper let denote the functor EC2+ ∧C2 −. Theorem 7.20. Let M be a discrete monoid with involution. There is an isomorphism of graded k-modules HO⋆ k[M] =∼ H⋆ B (C∞E, C1, M) .

Proof. Recall the C2-equivariant James construction J from Deﬁnition 7.6. By [Dun89, Lemma 2.5] there is a C2-equivariant weak homotopy equivalence B (J, J, M) → M, where B (J, J, M) is the geometric realization of the based simplicial C2-space B⋆ (J, J, M) given by the two-sided bar construction. We therefore have isomorphisms ∼ H ∼ H HO⋆ k[M] = H⋆ (hocolim M ) = H⋆ hocolim B(J,J,M) . By [May72, Lemma 9.7] we can take the hyperoctahedral bar construction and the homotopy colimit inside the bar construction and apply Corollary 7.18 to obtain, H ∼ H E hocolim B(J,J,M) = B hocolim J(−), J, M ≃ B (Z , J, M) . Finally, there is a weak homotopy equivalence B (ZE, J, M) ≃ B (C∞E, C1, M) by Corollary 7.18 and Proposition 7.8.

8. Hyperoctahedral Homology of Group Algebras We begin this section by demonstrating that hyperoctahedral homology of a group algebra, where the involution is induced from sending a group element to its inverse, is the homology of an infinite loop space on the two-sided bar construction. We will then show that in the case where this involution is a free C2-action away from the basepoint one can say more; the hyperoctahedral homology in this case is the homology of the C2-fixed points under the involution of an equivariant infinite loop space. In particular, this holds for all discrete groups of odd order. ∞ ∞ Let Q denote the free infinite loop space functor Ω Σ : Top⋆ → Top⋆. Proposition 8.1. Let G be a discrete group. Let the group algebra k[G] have involution determined by sending a group element to its inverse. There is an isomorphism of graded k-modules HO⋆ k[G] =∼ H⋆ QB (E, C1, G) .

Proof. Since G is a group, B (C∞E, C1, G) is group-like. We can therefore apply the version of the recognition principle given in [May77, VII, Theorem 3.1(ii)] to deduce a weak homotopy equivalence B (C∞E, C1, G) ≃ B (QE, C1, G). By [May72, Lemma 9.7] there is an isomorphism B (QE, C1, G) =∼ QB (E, C1, G). The proposition then follows from Theorem 7.20. 18 Let Φ: C2-Top⋆ → Top⋆ denote the C2-ﬁxed point functor.

Let ΣC2 : C2-Top⋆ → C2-Top⋆ denote the C2-equivariant suspension functor. For X ∈ 1 C2-Top⋆, the space ΣC2 (X)= S ∧ X has involution deﬁned by (t,x) = (1 − t, x) .

Let ΩC2 : C2-Top⋆ → C2-Top⋆ denote the C2-equivariant loop functor. For X ∈ C2-Top⋆, 1 the space ΩC2 (X) = HomTop⋆ S , X has involution deﬁned by f(t)= f(1 − t).

Let QC2 : C2-Top⋆ → C2-Top⋆ denote the C2-equivariant free inﬁnite loop space functor. That is, Q =Ω∞ Σ∞ . C2 C2 C2

Remark 8.2. Similarly to the non-equivariant case there is a natural C2-weak homotopy equivalence ΩC2 QC2 ΣC2 (X) ≃ QC2 (X).

Proposition 8.3. Let X ∈ C2-Top⋆ have a free C2-action away from the basepoint. There is E a natural weak homotopy equivalence of topological spaces Φ QC2 (X) ≃ Q (X).

Proof. A theorem of tom Dieck [Car92, Corollary 2.3] tells us that for a based C2-space X there E is a weak homotopy equivalence Φ QC2 (X) ≃ Q (X)×Q Φ(X) . Since we are assuming that the C -action on X is free, the fixed points Φ(X) are trivial and so Q Φ(X) is trivial. We 2 therefore deduce the required weak homotopy equivalence. In the non-equivariant case, Fiedorowicz [Fie84, Corollary 7.9] showed that for any topological monoid M there is a weak homotopy equivalence B (Σ, C1, M) → BM, between the bar construction and the classifying space of M. We will show that this extends to a C2-equivariant weak homotopy equivalence when M is a topological monoid with involution. Definition 8.4. For a topological monoid with involution M, we define an involution on C 1 Cn th Bn ΣC2 , 1, M = S ∧ 1 M, the n level of the bar construction, by (t,x) = (1 − t, x) where t ∈ S1 and x ∈ CnM. We define the geometric realization B Σ , C , M to have the 1 C2 1 involution induced by the level-wise involution. Definition 8.5. Let M be a topological monoid with involution. We define an involution on the classifying space BM by (t0,...,tn) , (m1,...,mn) = (tn,...,t0) , (mn,..., m1) . Remark 8.6. Let G be a discrete group with involution given by sending an element to its inverse. The involution on G induces an involution on the classifying space BG which is equivariantly homotopic to the involution of Definition 8.5 by [BF84, Construction 3.1].

Proposition 8.7. Let M be a topological monoid with involution. There exists a C2-weak C homotopy equivalence B QC2 , 1, M → ΩC2 QC2 BM. Proof. A routine check shows that the chain of equivalences constructed in [Fie84, Theorems 7.3 and 7.8] are equivariant with respect to the actions defined in [Ryb91, Section 1], yielding C a weak C2-equivariant weak homotopy equivalence B ΣC2 , 1, M → BM, which sends the involution of Definition 8.4 to the involution of Definition 8.5. Apply the functor ΩC2 QC2 to both sides of the equivalence. By [May72, Lemma 9.7] we can take these functors inside the bar construction on the left hand side. The result then follows from Remark 8.2. Theorem 8.8. Let G be a discrete group of odd order. Let the group algebra k[G] and the classifying space BG have the involution induced by sending a group element to its inverse. ∼ There is an isomorphism of graded k-modules HO⋆ k[G] = H⋆ ΦΩC2 QC2 BG .

Proof. By Theorem 7.20 we have an isomorphism HO ⋆ k[G] =∼H⋆ B (C∞E, C1, G) . Since G is a group, B (C∞E, C , G) is group-like and so by [May77, VII, Theorem 3.1(ii)] there is a weak 1 homotopy equivalence B (C∞E, C1, G) ≃ B (QE, C1, G). Since G has odd order, the involution is free away from the basepoint and we can apply Proposition 8.3, followed by [May72, Lemma 9.7] and Proposition 8.7 to obtain the following chain of equivalences E C C ∼ C B (Q , 1, G) ≃ B ΦQC2 , 1, G = ΦB QC2 , 1, G ≃ ΦΩC2 QC2 BG from which the theorem follows. 19 References [Aul10] Shaun V. Ault. Symmetric homology of algebras. Algebr. Geom. Topol., 10(4):2343–2408, 2010. [BE74] M. G. Barratt and Peter J. Eccles. Γ+-structures. I. A free group functor for stable homotopy theory. Topology, 13:25–45, 1974. [BF84] D. Burghelea and Z. Fiedorowicz. Hermitian algebraic K-theory of topological spaces. In Algebraic K- theory, number theory, geometry and analysis (Bielefeld, 1982), volume 1046 of Lecture Notes in Math., pages 32–46. Springer, Berlin, 1984. [BF04] Clemens Berger and Benoit Fresse. Combinatorial operad actions on cochains. Math. Proc. Cambridge Philos. Soc., 137(1):135–174, 2004. [BK72] A. K. Bousfield and D. M. Kan. Homotopy limits, completions and localizations. Lecture Notes in Mathematics, Vol. 304. Springer-Verlag, Berlin-New York, 1972. [Bro82] Kenneth S. Brown. Cohomology of groups, volume 87 of Graduate Texts in Mathematics. Springer- Verlag, New York-Berlin, 1982. [BV68] J. M. Boardman and R. M. Vogt. Homotopy-everything H-spaces. Bull. Amer. Math. Soc., 74:1117– 1122, 1968. [Car92] Gunnar Carlsson. A survey of equivariant stable homotopy theory. Topology, 31(1):1–27, 1992. [Dol62] Albrecht Dold. Universelle Koeffizienten. Math. Z., 80:63–88, 1962. [Dun89] Gerald Dunn. Dihedral and quaternionic homology and mapping spaces. K-Theory, 3(2):141–161, 1989. [Fie] Z. Fiedorowicz. The symmetric bar construction. URL: https://people.math.osu.edu/fiedorowicz.1/. [Fie84] Z. Fiedorowicz. Classifying spaces of topological monoids and categories. Amer. J. Math., 106(2):301– 350, 1984. [FL91] Zbigniew Fiedorowicz and Jean-Louis Loday. Crossed simplicial groups and their associated homology. Trans. Amer. Math. Soc., 326(1):57–87, 1991. [FT87] B. L. Fe˘ıgin and B. L. Tsygan. Additive K-theory. In K-theory, arithmetic and geometry (Moscow, 1984–1986), volume 1289 of Lecture Notes in Math., pages 67–209. Springer, Berlin, 1987. [Gra19] Daniel Graves. Functor Homology for Augmented Algebras, 2019. PhD Thesis. University of Sheffield. Available at http://etheses.whiterose.ac.uk/26243/. [Gra20] Daniel Graves. PROPs for involutive monoids and involutive bimonoids. Theory Appl. Categ., 35:No. 42, 1564–1575, 2020. [GZ67] P. Gabriel and M. Zisman. Calculus of fractions and homotopy theory. Ergebnisse der Mathematik und ihrer Grenzgebiete, Band 35. Springer-Verlag New York, Inc., New York, 1967. [Jam55] I. M. James. Reduced product spaces. Ann. of Math. (2), 62:170–197, 1955. [Joh90] D. L. Johnson. Presentations of groups, volume 15 of London Mathematical Society Student Texts. Cambridge University Press, Cambridge, 1990. [KLS87] R. L. Krasauskas, S. V. Lapin, and Yu. P. Solov’ev. Dihedral homology and cohomology. Basic concepts and constructions. Mat. Sb. (N.S.), 133(175)(1):25–48, 143, 1987. [Lod87] Jean-Louis Loday. Homologies diédrale et quaternionique. Adv. in Math., 66(2):119–148, 1987. [Lod90] Gerald M. Lodder. Dihedral homology and the free loop space. Proc. London Math. Soc. (3), 60(1):201– 224, 1990. [Lod98] Jean-Louis Loday. Cyclic homology, volume 301 of Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Springer-Verlag, Berlin, second edition, 1998. Ap- pendix E by Mar´ıa O. Ronco, Chapter 13 by the author in collaboration with Teimuraz Pirashvili. [LV12] Jean-Louis Loday and Bruno Vallette. Algebraic operads, volume 346 of Grundlehren der Mathematis- chen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Springer, Heidelberg, 2012. [May72] J. P. May. The geometry of iterated loop spaces. Springer-Verlag, Berlin-New York, 1972. Lectures Notes in Mathematics, Vol. 271. [May77] J. Peter May. E∞ ring spaces and E∞ ring spectra. Lecture Notes in Mathematics, Vol. 577. Springer- Verlag, Berlin-New York, 1977. With contributions by Frank Quinn, Nigel Ray, and Jørgen Tornehave. [McC01] John McCleary. A user’s guide to spectral sequences, volume 58 of Cambridge Studies in Advanced Mathematics. Cambridge University Press, Cambridge, second edition, 2001. [ML98] Saunders Mac Lane. Categories for the working mathematician, volume 5 of Graduate Texts in Mathe- matics. Springer-Verlag, New York, second edition, 1998. [PR02] T. Pirashvili and B. Richter. Hochschild and cyclic homology via functor homology. K-Theory, 25(1):39– 49, 2002. [Rot09] Joseph J. Rotman. An introduction to homological algebra. Universitext. Springer, New York, second edition, 2009. [Ryb91] S lawomir Rybicki. Z2-equivariant James construction. Bull. Polish Acad. Sci. Math., 39(1-2):83–90, 1991. [S l91] Jolanta S lomińska. Homotopy colimits on E-I-categories. In Algebraic topology Poznań1989, volume 1474 of Lecture Notes in Math., pages 273–294. Springer, Berlin, 1991. [Wei94] Charles A. Weibel. An introduction to homological algebra, volume 38 of Cambridge Studies in Advanced Mathematics. Cambridge University Press, Cambridge, 1994. 20 [Xic06] Miguel A. Xicoténcatl. On Z2-equivariant loop spaces. In Recent developments in algebraic topology, volume 407 of Contemp. Math., pages 183–191. Amer. Math. Soc., Providence, RI, 2006.